[Ffmpeg-devel-irc] ffmpeg.log.20130712

[burek]

[00:40] <juanmabc> hi, since recently someone work on adpcm ima radical, i wonder if some ffmpeg developer would like, or get interested, on work on adpcm ima amv encoder, else i'll keep waiting for another year until someone (if ever) does (before devices die would be better ;)), there is some info, first even a ticket time ago (known for me from ages): http://trac.ffmpeg.org/ticket/747 , second a working code to base: http://code.google.com/p/amv-codec-
[00:40] <juanmabc> tools/ and third a working, in trunk, adpcm ima amv decoder to test the results (cause not everyone got the gadget ;)), so the workflow is far from scratch, and far from hard to test. Discalimer: this is just a push, ;)
[00:41] <juanmabc> to summarize, adpcm ima amv enc, now
[00:41] <juanmabc> ;P
[04:31] <Lobs> is there a way i can force an output resolution in a mov container? I.E the encoded resolution differs from the playback resolution and isnt directly linked to the aspect ratio.
[05:26] <calcifer_> is an updated homebrew formula for ffmpeg 2.0 in the works?
[05:29] <llogan> calcifer_: i don't think anyone here is responsible for that
[05:29] <llogan> i mean i
[05:29] <llogan> 'm not sure who updates that (damned wrist brace)
[05:33] <calcifer_> looking at the history, it's been updated by a few different people. hmm. :/
[09:51] -:#ffmpeg- [freenode-info] why register and identify? your IRC nick is how people know you. http://freenode.net/faq.shtml#nicksetup
[10:30] <t4nk488> hi, how to mux h264 video and pcm_s16le by ffserver
[10:30] <t4nk488> I try to output a rtp stream that include both video and audio. the config I use:  <Stream test1-rtsp.mpg> Feed feed1.ffm Format rtp AVOptionVideo flags +global_header VideoSize 1280x720 VideoFrameRate 30 VideoCodec libx264 AVOptionAudio flags +global_header AudioCodec pcm_s16le AudioChannels 2 AudioSampleRate 48000 </Stream>
[10:31] <t4nk488> I can NOT play by ffplay and VLC.
[10:31] <t4nk488> Does anyone have idea how to mux h264 video and pcm_s16le audio?
[10:39] <Mavrik> hmm, I think the raw audio is the problem
[10:39] <Mavrik> why aren't you encoding it?
[11:06] <sweb> how can i strip all metatags from file with ffmpeg and then add some tags ...
[11:07] <sweb> ffmpeg -i 1.mp3 -map_metadata -1 2.mp3 will strip all tags. but how can i add some tags after then
[11:22] <t4nk488> hi
[11:22] <t4nk488> what is the error:ffserver: libavformat/asfenc.c: 642: flush_packet: Assertion `asf->packet_timestamp_end >= asf->packet_tim estamp_start' failed.
[11:22] <t4nk488> does anyon know the cause of the ffserver fail
[11:46] <t4nk488> hi, any ideas about how to mux h264 video + pcm audio?
[11:47] <t4nk488> Can anyone help?
[11:50] <Mavrik> mux to what.
[11:53] <t4nk488> mux to network stream
[11:53] <t4nk488> I try to mux asf stream, but VLC can not play
[11:54] <t4nk488> and I mux tp rtp stream , VLC can not play ,too
[11:54] <t4nk488> both VLC and ffplay can not play the stream
[11:54] <Mavrik> streaming raw anything is a problem
[11:54] <Mavrik> encode it to AAC and it'll work.
[11:56] <t4nk488> I also try to encode to AAC , but when I try to mux to rtp stream
[11:56] <t4nk488> VLC can not play,too
[11:57] <t4nk488> I got the error message:Application provided invalid, non monotonically increasing dts to muxer in stream
[11:57] <t4nk488> do you have idea about :Application provided invalid, non monotonically increasing dts to muxer in stream
[11:57] <t4nk488> hence VLC can not play asf stream
[11:59] <Mavrik> well if you have messed up timestamps
[11:59] <Mavrik> then no player will be able to play it no matter ne mux
[12:00] <t4nk488> do you have idea about tiemstamps issue?
[12:03] <t4nk488> do you have any idea: which encode mthod is better for audio other than aac?
[14:30] <FunnyLookinHat> Hey guys - So I've got some scripts that generate a bunch of thumbnail preview images from a video using something like this: ffmpeg -i input.mp4 -ss 45 -f image2 -vframes 1 out.png  ... but I've noticed it takes a significantly long time as the timestamp gets further out ( several minutes sometimes to get something just 3 or 4 minutes into the video )
[14:30] <FunnyLookinHat> Is there any way to do this that's faster using ffmpeg?
[14:33] <Mavrik> move -ss before -i.
[14:40] <FunnyLookinHat> Mavrik, That will seek faster ?
[14:41] <JEEB> yes
[14:41] <JEEB> it will, though, only be able to seek to points that it can start decoding from straight away
[14:42] <JEEB> seek on the container level (before decoding) vs seek on the stream level (with decoding)
[14:42] <JEEB> latter is more accurate, but will decode everything, first is faster
[14:43] <Mavrik> also on some files it might not work
[14:49] <FunnyLookinHat> Interesting...  ok I'll do a bunch of testing.  Much appreciated, thanks guys.
[14:56] <i0scribe> hello everyone.
[14:56] <i0scribe> I just followed the UbuntuCompilationGuide at the Wiki and wasn't sure if it "took" successfully.
[14:58] <i0scribe> At the end of the guide, is directed me to give the following command "ffmpeg 2>&1 | head -n1" and my output is "ffmpeg version 0.8.6-4:0.8.6-0ubuntu0.12.04.1, Copyright (c) 2000-2013 the Libav developers." however the guide (which I believe is current) says "If you do not see FFmpeg developers in your output then something went wrong and you're probably using the  fake "ffmpeg" from the repository.".
[15:00] <i0scribe> So, my question is, what is the difference between "Libav developers" and "FFmpeg developers" when it comes to knowing if my compilation was completed successfully.
[15:03] <JEEB> libav and ffmpeg are separate projects
[15:03] <JEEB> debian-based distros use binaries built from the libav project's source code by default
[15:03] <JEEB> so if you see "ffmpeg developers" there, it means you are indeed using a binary that is built using the ffmpeg project's source code
[15:19] <i0scribe> Thanks JEEB ... So I would see "FFmpeg developers" if I was using Redhat (or some other distro) and "Libav developers" is just the expected result on all Ubuntu and other deb based distros ... ok, thanks.
[15:25] <i0scribe> Since I've confirmed that my ffmpeg is current and unrestricted, can anyone offer me any tips as to why my AVI to MPG conversions aren't going well? I've tried three good working AVIs coming from two different sources, and each one gives me sound only. Well, that is except for one "still image" snapshot. I think the "still image" is the first image from the AVI file.
[15:26] <i0scribe> I'm using WinFF which does a fine job of converting FLV to MPG but the AVI to MPG is not working.
[15:44] <i0scribe> I tried "ffmpeg -i file.avi -vcodec mpeg4 -f mpeg output.mp4" and that is working. Also tried "ffmpeg -i file.avi -vcodec h263 -acodec aac -f mpg output.mpg" ... Am looking for some other options on the net hoping to get this AVI converted successfully to an MPG.
[15:47] <Mavrik> uh
[15:47] <Mavrik> ok.
[15:47] <Mavrik> that I can see why it's going wrong >D
[15:47] <Mavrik> i0scribe, what are you trying to achieve? why (the horror) MPG?
[15:49] <i0scribe> Maverick|MSG: Trying to convert it to MPG because it's the only thing supported by my DVD player (USB port).
[15:49] <Maverick|MSG> Mavrik ^^
[15:49] <Mavrik> ;)
[15:49] <Mavrik> i0scribe, well, MPG is just the container
[15:50] <Mavrik> first you MUST check which video and audio formats your DVD player supports
[15:50] <Mavrik> and then we can talk
[15:50] <Mavrik> stuffing random stuff into MPG just won't work ;)
[15:50] <i0scribe> I also tried "ffmpeg.exe -i input.avi output.mpg" just now and that's not working.
[15:50] <ilove11ven> i0scribe: you may want to use something like dvd author, which use FFmpeg internally to do it.
[15:51] <i0scribe> Well, for now I'm just trying to get the MPG file to play on my computer and I can't even get that far.
[15:51] <i0scribe> Trying to convert a working AVI file to a working MPG file.
[15:51] <ilove11ven> i0scribe: Mavrik is correct. To rip DVD, not only did you have to use -f dvd, but you should also take into account of codec problem and many other complicated stuff.
[15:52] <ilove11ven> MPEG-2 video and MPEG audio could be used with '-f dvd'
[15:54] <Mavrik> i0scribe, which DVD player is it?
[15:54] <i0scribe> ilove11ven: I think I confused the issue by bringing up my DVD player ... My DVD player has a USB port where I can play normal MPG files via a USB stick. ... So nothing special is required on that side. ... I can take 'any' working MPG file that I pull off my computer and play it on my DVD player.
[15:55] <i0scribe> Mavrik: I've used both "Mobile Media Converter" and "WinFF" to successfully convert FLV files to MPG, but when that didn't work I decided to drop down to the CLI and see what was going on with FFmpeg.
[15:55] <Mavrik> dude.
[15:55] <Mavrik> mpg is just a container
[15:55] <Mavrik> you need to put a correct video and audio format in there.
[15:55] <Mavrik> that's why we're asking, WHICH DVD player do you have
[15:55] <Mavrik> so we can help you
[15:56] <Mavrik> running random crap won't just make it work except by blind luck
[15:59] <ilove11ven> i0scribe: try this: ffmpeg -i .avi -f dvd -vcodec mpeg2video -acodec mp2 output.mpg
[16:02] <ilove11ven> or take any working mpg file, use FFtools to analyze its container/video codec/audio codec. Then use ffmpeg to do the conversion.
[16:03] <i0scribe> Maverick|MSG: It's a Xenon DVD player.
[16:04] <i0scribe> ilove11ven: thanks, I'll try that.
[16:06] <i0scribe> ilove11ven: oh wait, what part of that string is the input.avi ?
[16:06] <i0scribe> ilove11ven: I'm guessing to ".avi" part of it.
[16:06] <ilove11ven> your avi file
[16:07] <Mista_D> Anyway to force "-vf scale=-1:480" to use nearset disible by 16?
[16:08] <Mavrik> Mista_D, I think you can use expressions
[16:10] <Mavrik> so you could do something like
[16:10] <Mavrik> scale=round((oh * a) / 16) *16:480
[16:10] <Mavrik> or something like taht
[16:10] <Mavrik> :)
[16:10] <i0scribe> ilove11ven: same problem as before. Audio only as the output. The 28 MB AVI file always comes out as a half meg MPG file that is audio only.
[16:11] <i0scribe> ilove11ven: Don't know if this is useful for you, but every time I do this one of the error lines says "[buffer @ 0x87a99c0] Changing frame properties on the fly is not supported.".
[16:11] <Mavrik> Mista_D, I doubt there's a simpler ffmpeg-only way
[16:12] <Mavrik> sorry, it's floor() not round()
[16:12] <Mista_D> Mavrik: Doesn't work - "Unknown function", using 5 hr old master.
[16:12] <Mavrik> https://ffmpeg.org/ffmpeg-utils.html#Expression-Evaluation
[16:12] <Mavrik> writing out of my head ;)
[16:13] <Mavrik> anyway, bbl
[16:13] <Mista_D> Mavrik: Will test a few, thanks.
[16:15] <ilove11v1> i0scribe: would you mind uploading the avi file?
[16:15] <i0scribe> no prob.
[16:15] <i0scribe> ilove11v1: no prob.
[16:18] <ilove11v1> ok
[16:20] <i0scribe> ilove11v1: just started the file transfer.
[16:21] <i0scribe> ilove11v1: it's says it is waiting for you.
[16:54] <Mista_D> Why would this scale folmula be rejected: -vf 'scale=round((iw*480/ih)/16) *16:480'? seems ok to me.
[17:00] <saste> Mista_D, what is "round"?
[17:01] <Mista_D> @saste: Tried celi too.
[17:01] <Mista_D> Sorry ceil
[17:03] <Mista_D> @saste: stand corrected,  -vf 'scale=ceil((iw*480/ih)/16) *16:480' works. Had a differnt error, didn't noticed at first.
[17:04] <ilove11v1> i0scribe have upload the original file: http://www.fileconvoy.com/dfl.php?id=g0a0c8b0c8b7f9a9d999327977b1aa086cf5a5fe82
[17:05] <ilove11v1> I have uploaded the transcoded MPEG-PS/MPEG2 video/MP2 audio: http://www.fileconvoy.com/dfl.php?id=ge1604f2b7b8f9e6b999327998151be79b266fd98c
[17:05] <ilove11v1> i0scribe: please check it. time to go to bed. good night
[19:29] <teratorn> what aac codec to folks normally recommend these days?
[19:30] <Mavrik> fdk_aac is probably still the best
[19:30] <Mavrik> otherwise internal is good enough
[19:30] <teratorn> cool. now to figure out how to use it from xuggle :-(
[19:30] <Fjorgynn_> Mavrik: I heard different
[19:30] <Fjorgynn_> that it's not good. The default
[19:30] <teratorn> or that. we're getting sporadic SIGFPE exceptions from the lovely libvo-aacenc that xuggle uses by default
[19:30] <Mavrik> Fjorgynn_, none of the others are any better.
[19:31] <Mavrik> teratorn, yeah, voaacenc is probably the worst of the bunch
[19:35] <JEEB> <teratorn> what aac codec to folks normally recommend these days? <- do you mean encoder or decoder wise?
[19:36] <teratorn> encoder
[19:36] <JEEB> ok, then fdk-aac
[19:36] <JEEB> do note that you will have to compile fdk-aac yourself, it can't be distribute
[19:36] <JEEB> *distributed
[19:36] <JEEB> as binary
[19:36] <teratorn> ah
[19:36] <Fjorgynn_> I have the binary for windows :)
[19:37] <JEEB> yes, but according to the license you cannot distribute it :)
[19:38] <JEEB> VideoLAN tried to find some ways around it
[19:38] <JEEB> but in the end just gave up
[19:39] <teratorn> what is the license? according to the .c file it says "is part of FFMPeg... FFMpeg is distributed under LGPL blah blah"
[19:41] <JEEB> teratorn, that's the license for the wrapper in ffmpeg
[19:41] <JEEB> teratorn, https://github.com/mstorsjo/fdk-aac/blob/master/NOTICE
[19:41] <JEEB> this is the license for fdk-aac
[19:41] <teratorn> yeah I just realized 411 lines can't be enough for the full codec
[19:41] <teratorn> cool, thanks
[19:53] <eggyknap> If ffprobe tells me "Stream #0:1(und): Audio: aac (mp4a / 0x6134706D), 48000 Hz, 4.0, fltp, 51 kb/s", the "4.0" means I have four channels of audio in one stream, right?
[19:54] <JEEB> I would think so, yes
[19:54] <eggyknap> (follow-up questions: how can I extract one of those channels into a separate mp3, and second, I created that file with ffmpeg and would really like one stereo and two mono streams... how can I do that?)
[19:56] <JEEB> there are IIRC options to use certain channels and such
[19:56] <JEEB> I don't remember them all myself, but looking at the documentation for the ffmpeg command line tool on ffmpeg.org should be useful
[19:56] <eggyknap> Yeah, I've been trying to find examples, because I haven't been able to make those options work (disclaimer: the ffmpeg version I've got on my ubuntu box seems hopelessly disabled -- old version, with few enabled filters)
[19:57] <eggyknap> IOW, I've found the documentation, and have read about the options, but am getting something wrong.
[19:57] <JEEB> with ubuntu you will most probably want to use the avconv binary as you're using code from the libav project
[19:57] <JEEB> in libav avconv is the more up-to-date binary
[19:58] <JEEB> while in ffmpeg (project), ffmpeg is
[19:58] <JEEB> (also #libav is the IRC channel for the libav project)
[19:58] <eggyknap> I had the impression (based only on manpages and my inept blundering around) that on my box they were the same binary, essentially. I'll look more closely into it.
[19:58] <eggyknap> In any case, my avconv doesn't support -filter_complex any better than my ffmpeg does
[19:59] <JEEB> I know that at least until version 0.8 of libav there was still a separate ffmpeg binary, I think?
[19:59] <JEEB> after the 0.8 release it got finally removed
[19:59] <JEEB> and the ffmpeg binary, while it was there, was pretty much untouched
[19:59] <JEEB> also if you're looking for the avconv documentation, I recommend you look at libav.org's documentation :)
[20:01] <eggyknap> So, say I have a camera and  multiple microphones, and I say something like "ffmpeg <options to capture video> <options to capture mic 1> <options to capture mic 2> outputfile.mp4"...
[20:01] <eggyknap> ...I'd like to end up with multiple audio tracks, rather than multiple channels on one track. Suggestions?
[20:02] <Mavrik> just map them accordingly to different tracks and set encoding options to each
[20:03] <eggyknap> Mavrik: while that sentence makes sense, I'm too much of a n00b to see how to do that :)
[20:04] <Mavrik> I suggest reading the documentation then
[20:04] <Mavrik> you can provide multiple inputs (one -i for each input) and then use -map parameter to map inputs to a single output file with different streams
[20:05] <eggyknap> Clearly I need to grok -map better than I do. I'll reread its docs. Thx
[20:05] <Mavrik> you provide encoding parameter with each stream as well
[20:05] <Mavrik> and you give only one output
[20:05] <Mavrik> in the docs it's explained how to use the -map parameter and how to number encoding parameters (codec:a, codec:v etc.) for each output stream :)
[20:06] <Mavrik> eggyknap, also by default each input will be a separate stream
[20:06] <Mavrik> so if I read the docs correctly you shouldn't need a -map call ata all
[20:07] <eggyknap> Hrm. I didn't include any options to my capture command and ended up with two streams, one in 4.0 audio
[20:08] <Mavrik> does your audio input record 4 channels? :)
[20:08] <Mavrik> eggyknap, maybe you should pass parameter for only 1 output channel for that stream then :)
[20:08] <Mavrik> especially if your microphone records mono
[20:08] <Mavrik> eggyknap, it's worth checking what ffmpeg thinks the INPUT is
[20:10] <eggyknap> There are four separate mono mics, if that's what you mean. How do I get ffmpeg to cough up more data on what it thinks its INPUT is?
[20:10] <Fjorgynn> soon it's the 27th...
[20:11] <Mavrik> eggyknap, um, it always outputs what it thinks the input is
[20:13] <eggyknap> Oh, I just need to pay more attention to the STDOUT stuff. Gotcha :)
[21:09] <llogan> eggyknap: try -map_channel (or pan audio filter) to monkey with the 4 audio channels
[21:10] <eggyknap> Ooo... that could be helpful
[21:10] <eggyknap> Right now I'm recording from four mics simultaneously, and the result I get is a file with one mono output stream. :(
[21:13] <eggyknap> e.g. http://pastie.org/8135391
[21:17] <eggyknap> I can add "-filter_complex '[0:0] [1:0] [2:0] [3:0] amerge=inputs=4'" and get a single audio stream in 4.0 layout... but that's still not four separate streams.
[21:17] Action: eggyknap keeps reading
[21:18] <eggyknap> OIC... there's where I need a -map command.
[21:18] <eggyknap> ...perhaps
[21:19] <eggyknap> w00t! That works. Thanks, all.
[22:12] <nuizzy> does ffmpeg support h.265?
[22:14] <Sean__> I want to analyze a source video file and determine 1. At What frames each  "scene change" occurs.  2. Given the information from #1 and a min and max key frame interval , what frames will need to be I-frames in a h264 encode..
[22:14] <Sean__>  I know I can do this by doing a "ultrafast" pass=1 encode with x264 and analyzing the .stats file afterwards.  However , This method seems a bit heavy for the simple information I'm trying to obtain
[00:00] --- Sat Jul 13 2013